A Genome-Scale DNA Repair RNAi Screen Identifies
SPG48 as a Novel Gene Associated with Hereditary
Mikołaj Słabicki1, Mirko Theis1, Dragomir B. Krastev1, Sergey Samsonov2, Emeline Mundwiller3,4,5,
Magno Junqueira1, Maciej Paszkowski-Rogacz1, Joan Teyra2, Anne-Kristin Heninger1, Ina Poser1,
Fabienne Prieur6, Je ´re ´my Truchetto3,4,5, Christian Confavreux7, Ce ´cilia Marelli3,4,5,8, Alexandra
Durr3,4,5,8, Jean Philippe Camdessanche6, Alexis Brice3,4,5,8, Andrej Shevchenko1, M. Teresa Pisabarro2,
Giovanni Stevanin3,4,5,8, Frank Buchholz1*
1MaxPlanckInstituteforMolecularCell BiologyandGenetics,Dresden,Germany,2Structural Bioinformatics,BIOTECTU,Dresden,Germany,3INSERM,Unit975Paris,France,
4Universite ´ Pierre et Marie Curie-Paris6, Centre de Recherche de l’Institut du Cerveau et de la Moelle Epinie `re, Paris, France, 5CNRS, Unite ´ Mixte de Recherche 7225 Paris,
France,6Ho ˆpitalNord,Saint Etienne,France,7Ho ˆpital Neurologique,Lyon, France,8APHP,Pitie ´-Salpe ˆtrie `reHospital,DepartmentofGeneticsandCytogenetics,Paris,France
DNA repair is essential to maintain genome integrity, and genes with roles in DNA repair are frequently mutated in a variety
of human diseases. Repair via homologous recombination typically restores the original DNA sequence without introducing
mutations, and a number of genes that are required for homologous recombination DNA double-strand break repair (HR-
DSBR) have been identified. However, a systematic analysis of this important DNA repair pathway in mammalian cells has
not been reported. Here, we describe a genome-scale endoribonuclease-prepared short interfering RNA (esiRNA) screen for
genes involved in DNA double strand break repair. We report 61 genes that influenced the frequency of HR-DSBR and
characterize in detail one of the genes that decreased the frequency of HR-DSBR. We show that the gene KIAA0415 encodes
a putative helicase that interacts with SPG11 and SPG15, two proteins mutated in hereditary spastic paraplegia (HSP). We
identify mutations in HSP patients, discovering KIAA0415/SPG48 as a novel HSP-associated gene, and show that a
KIAA0415/SPG48 mutant cell line is more sensitive to DNA damaging drugs. We present the first genome-scale survey of HR-
DSBR in mammalian cells providing a dataset that should accelerate the discovery of novel genes with roles in DNA repair
and associated medical conditions. The discovery that proteins forming a novel protein complex are required for efficient
HR-DSBR and are mutated in patients suffering from HSP suggests a link between HSP and DNA repair.
Citation: Słabicki M, Theis M, Krastev DB, Samsonov S, Mundwiller E, et al. (2010) A Genome-Scale DNA Repair RNAi Screen Identifies SPG48 as a Novel Gene
Associated with Hereditary Spastic Paraplegia. PLoS Biol 8(6): e1000408. doi:10.1371/journal.pbio.1000408
Academic Editor: Nicholas Hastie, Medical Research Council Human Genetics Unit, United Kingdom
Received March 12, 2010; Accepted May 19, 2010; Published June 29, 2010
Copyright: ? 2010 Słabicki et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study was funded by the Max Planck Society (to FB, AS); by the German Federal Ministry of Education and Research grants Go-Bio  (to
FB), NGFN-plus [01GS0859] (to FB); by the Klaus Tschira Stiftung GmbH (to JT, SS, and MTP); by NIH NIGMS grant 1R01GM070986-01A1 (to AS); by the European
Union grant EUROSPA consortium (to AB); by the French National Agency for Research grants ANR-SPG11 (to GS) and ANR-SPAX (to AD); and by the Verum
foundation (to AB). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
Abbreviations: BAC, bacterial artificial chromosome; BER, base excision repair; PDB, Brookhaven Protein Databank; DSB, double strand break; HSP, hereditary
spastic paraplegia; IR, ionizing radiation; LOH, loss of heterozygosity; MMC, mitomycin C; NER, nucleotide excision repair; NHEJ, nonhomologous end joining; PFA,
paraformaldehyde; Rluc, renilla luciferase; TCC, thin corpus callosum
* E-mail: firstname.lastname@example.org
Mutations in DNA repair genes are associated with different
diseases and disorders including cancer , accelerated aging ,
and neuronal degeneration . Neurons appear to be particularly
vulnerable to mutations in DNA repair genes, possibly due to the
lack of proliferation and high oxidative stress within these cells. As
a consequence, several neurological diseases have been linked to
defects in DNA repair such as Ataxia-telangiectasia , Ataxia-
telangiectasia-like disorder , Seckel syndrome , Nijmegen
breakage syndrome , and Charcot-Marie-Tooth syndrome .
A particularly dangerous DNA lesion for a cell is a double
strand break (DSB), in which two strands of the DNA are broken
in close proximity to one another [9,10]. DSBs are repaired
mainly via two parallel pathways: homologous recombination and
nonhomologous end joining (NHEJ). Repair via homologous
recombination typically restores the genetic information, whereas
repair via NHEJ often leads to mutations [10,11].
Recently, several RNAi screens have addressed different aspects
of mammalian DNA repair, such as increased sensitivity towards
PARP inhibition , increased sensitivity towards cisplatin ,
accumulation of 53BP1 foci [14,15], or altered phosphorylation of
the histone variant H2AX . These screens have greatly
enhanced our understanding of human DNA repair processes
and delivered a number of novel genes implicated in various
aspects of DNA repair. Here, we report a genome-scale RNAi
screen for genes implicated in homologous recombination-
mediated DSB repair, uncovering a variety of known and so far
PLoS Biology | www.plosbiology.org1 June 2010 | Volume 8 | Issue 6 | e1000408
uncharacterized genes implicated in this process. In this work, we
mine this screen employing a structural bioinformatics approach
and identify KIAA0415/SPG48 as a putative helicase that is
associated with hereditary spastic paraplegia (HSP).
Genome-Scale RNAi Screen
For a comprehensive search of genes associated with DNA DSB
repair, we performed a genome-scale RNAi screen, utilizing an
endoribonuclease-prepared short interfering RNA (esiRNA) li-
brary  and employing the well-established DR-GFP assay
. First, a stable HeLa cell line with two non-functional GFP
alleles was generated, in which GFP expression is efficiently
activated only after HR-DSBR (Figure 1A). We then tested the
robustness of the assay by co-transfection of these cells with the I-
SceI expression plasmid and an esiRNA targeting Rad51, which is
an essential factor for the early stages of homologous pairing and
strand exchange . Depletion of Rad51 resulted in a marked
reduction of GFP positive cells, and comparisons to negative
control transfected cells suggested a high dynamic range for
candidate factors influencing HR-DSBR (Figure 1B and histo-
grams Figure 1C).
TheRNAiscreen wascarried outinduplicatein384-wellplatesby
co-transfection of an I-SceI encoding plasmid with the individual
esiRNAs targeting over 16,000 human genes . The percentage of
GFP positive cells was determined by high throughput FACS,
providing a sensitive readout for esiRNAs influencing the frequency
of HR-DSBR (Figure 1C). Knockdown of 228 and 141 transcripts
significantly decreased or increased the percentage of GFP positive
cells, respectively (Figure 1D, Table S1). Among the strongest
knockdowns affecting HR-DSBR were genes with well-characterized
roles in DNA repair such as Rad51, BRCA1, and SHFM1. Gene
ontology enrichment analysis of the candidates revealed a 5-fold
enrichment for genes reported to be implicated in DNA repair
(Figure 1E), confirming that the screen was efficient.
To validate the candidate hits we examined their expression in
HeLa cells and resynthesized all esiRNAs for the genes that were
expressed. We also generated a second, independent, and non-
overlapping esiRNA for these genes and tested all esiRNAs again
in the DR-GFP assay in multiple replicates. Using stringent
selection criteria (see Online Methods), 45 genes decreased the
frequency of homologous recombination, while 17 genes increased
it with two independent silencing triggers (Table 1). To further
narrow down the list of these 62 candidates, we tested the esiRNAs
for their impact on intracellular GFP levels. EsiRNAs that
influence GFP levels, for example by targeting a transcriptional
activator for GFP expression, could score in the DR-GFP assay
and contaminate the hit list. We therefore transfected the esiRNAs
into GFP expressing HeLa cells and assayed GFP levels by FACS.
EsiRNAs targeting MKNK2 reduced GFP levels in these cells.
Therefore, this gene was excluded from further analysis, reducing
the final hit list to 61 genes (Table 1). The effectiveness of this
stringent validation was monitored again by gene ontology
enrichment analysis, with an enrichment of now 20-fold for genes
annotated in the category DNA repair (Figure 1E).
Knockdowns That Increased the Frequency of HR-DSBR
Silencing of 17 genes significantly increased the number of GFP
positive cells in the DR-GFP assay. Hence, the knockdown of these
genes promoted HR-DSBR, which might be of interest for several
biological applications such as increasing the targeting efficiency of
genes by homologous recombination . Different reasons might
account for the increased number of GFP positive cells observed.
One possibility is that the knockdown led to an inhibition of the
NHEJ pathway, thereby shifting the ratio of the two possible
pathways toward repair via HR. Support for this reasoning comes
from experiments in yeast and flies, where the knockout of DNA
ligase IV, a gene that is required for NHEJ , significantly
increased gene targeting by homologous recombination [21,22].
Interestingly, the knockdown of human Lig4 resulted in a striking
increase in GFP positive cells in the DR-GFP assay (Table 1),
suggesting that inhibition of the NHEJ pathway can increase the
frequency of HR-DSBR also in mammalian cells. This idea is
further supported by inspection of other known NHEJ proteins,
including XRCC4, XRCC5, XRCC6, PRKDC, and DCLRE1C
[23,24]. Knockdown of all of these proteins increased the
frequency of homologous recombination in the DR-GFP assay
(Table S1). Hence, we speculate that other genes that increased
the number of GFP positive cells might be implicated in the NHEJ
pathway and that knockdown of these genes could enhance gene
targeting by homologous recombination in mammalian cells.
Knockdowns That Decreased the Frequency of HR-DSBR
The list of genes that decreased the frequency of HR-DSBR was
enriched for proteins with well-defined roles in HR-DSBR, such as
Rad51 and BRCA1. In addition, genes, such as E2F1, that more
indirectly influence HR-DSBR were also identified in the screen.
E2F1 is involved in cell cycle and apoptosis regulation after DNA
damage  and has recently been implicated in transcriptional
regulation of Rad51 and BRCA1 , possibly explaining why the
knockdown of E2F1 scored in our screen. Interestingly, the assay
also uncovered a number of genes that have roles in DNA repair
processes other than HR-DSBR, such as XPC, which has a role in
nucleotide excision repair (NER) , and the base excision repair
(BER) DNA helicase RECQL4 . However, a polymorphism in
the XPC gene has recently been shown to correlate with
bleomycin-induced chromosomal aberrations , and RECQL4
has been reported to coincide with foci formed by Rad51 after
induction of DSBs , suggesting possible links between the
different DNA repair pathways. Finally, the gene list is enriched
for proteasome subunits, including PSMD4, PSMD1, PSMD14,
and SHFM1. Treatment with proteasome inhibitors has been
shown to specifically suppress HR-DSBR possibly because of the
lack of proteasome-mediated degradation of chromatin bound
All cells in our bodies have to cope with numerous lesions
to their DNA. Cells use a battery of genes to repair DNA
and maintain genome integrity. Given the importance of
an intact genome, it is not surprising that genes with roles
in DNA repair are mutated in many human diseases. Here,
we present the results of a genome-scale DNA repair
screen in human cells and discover 61 genes that have a
potential role in this process. We studied in detail a
previously uncharacterized gene (KIAA0415/SPG48) and
demonstrated its importance for efficient DNA double
strand break repair. Further analyses revealed mutations in
the SPG48 gene in some patients with hereditary spastic
paraplegia (HSP). We showed that SPG48 physically
interacts with other HSP proteins and that patient cells
are sensitive to DNA damaging drugs. Our data suggest a
link between HSP and DNA repair and we propose that
HSP patients should be screened for KIAA0415/SPG48
mutations in the future.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org2 June 2010 | Volume 8 | Issue 6 | e1000408
Figure 1. Genome-scale HR-DSBR esiRNA screen. (A) Schematic representation of the DR-GFP assay. The two non-functional GFP alleles and the
I-SceI cutting site are shown. The transfected plasmid encoding the I-SceI endonuclease is presented as a red circle. The functional GFP gene that is
generated after successful HR-DSBR is shown in green. (B) Immunofluorescence analysis of the DR-GFP HeLa cell line after transfection with or
without the I-SceI endonuclease plasmid and indicated esiRNAs. Scale bars represent 10 mm. (C) Analysis of an example plate from the screen. Grey
wells indicate knockdowns that did not significantly change the percentage of GFP positive cells observed, and red and green wells denote
knockdowns that decreased or increased the percentages of GFP positive cells observed, respectively. Control wells are marked with black frames. On
each plate there were four positive controls (esiRNA targeting Rad51) and eight negative controls (esiRNA against Rluc - renilla luciferase). Example
FACS histograms for the control transfections are presented. (D) Dot plot of the primary screen. Results are presented as average z-scores derived
from two independent replicates. Knockdowns with z-scores below 22 or above 2 are shown in red or green, respectively. (E) Results of the gene
ontology enrichment analysis for the primary (black) and validated (grey) hits.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org3 June 2010 | Volume 8 | Issue 6 | e1000408
Table 1. Summary of phenotypic data of the candidate genes implicated in HR-DSBR.
Gene NameEnsembl ID
Frequency of HR-DSBR
GFP Levels Viability
esiRNA 1 esiRNA 2 CisplatinMMCIR
Q 1h, Q 6h
n.d n.dn.d n.d
n.d n.d n.d n.d
n.dn.d n.d n.d
n.dn.d n.d n.d
n.dn.d n.d n.d
n.d n.dn.d n.d
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org4June 2010 | Volume 8 | Issue 6 | e1000408
proteins blocking the access to the lesion [31,32]. Moreover,
SHFM1 has been shown to be required for Rad51 foci formation
upon DNA damage , implicating a more direct role of this
proteasome subunit in HR-DSBR and possibly providing an
explanation why SHFM1 was one of the strongest hits in our
screen. Based on these results we were encouraged to investigate
further the knockdowns that decreased the number of GFP
positive cells in the DR-GFP assay.
To characterize in detail the 44 knockdowns that decreased the
frequency of HR-DSBR, we performed several additional assays.
First, we tested the influence on cell viability of these esiRNAs in
HeLa cells. Thirteen esiRNAs considerably decreased cell
numbers and were excluded from further analyses (Table 1).
Second, we performed mitomycin C (MMC), cisplatin, and
ionizing radiation (IR) sensitivity assays. MMC predominantly
causes interstrand cross-links, which result, among other things, in
DSBs due to a block of replication forks . Cisplatin damages
DNA in a different way and generates predominantly intrastrand
cross-links , whereas IR gives rise to a variety of DNA lesions
. Cells with impaired DNA repair pathways might be more
sensitive to these treatments, which should manifest in reduced cell
viability. Twenty-four hours post-transfection of the esiRNAs, the
cells were treated for 1 h with MMC, cisplatin, or exposed to IR
and cells were counted after an additional incubation for 48 h. A
number of knockdowns increased the sensitivity towards one or
more treatments, substantiating a role of these genes in DNA
repair, with some of the knockdowns showing an effect for one, but
not the other treatment (Figure 2A, Table 1). For instance, the
knockdown of RBBP8 (also known as CtIP), which promotes DNA
end resection , did not cause increased sensitivity towards
after MMC treatment, indicating that RBBP8 depletion primarily
sensitized the cells against this drug. Third, we employed
a gamma-H2AX removalassay.
less cellswere counted
phosphorylated on serine 139 predominantly by ATM/ATR
[38,39] at sites of DSBs until the lesion is repaired. After successful
DNA repair this phosphorylation is reverted by the phosphatase
PP2A . Several knockdowns resulted in extended time before
gamma-H2AX was removed from irradiated cells (Figure 2B,
Table 1), suggesting a delay in DSBR, and potentially explaining
the observed reduction of GFP positive cells in the DR-GFP assay.
Surprisingly, a few knockdowns showed overall reduced numbers
of gamma-H2AX positive cells, or accelerated removal of gamma-
H2AX after irradiation. For example, depletion of ARHGEF1
resulted in a reduced number of gamma-H2AX positive cells 1 h
after irradiation (Figure 2B). Potentially, this Rho guanine
nucleotide exchange factor  is required for efficient recruit-
ment of H2AX phosphorylation factors, which ultimately
translates into less efficient HR-DSBR. In contrast, the knockdown
of FIZ1, a Flt3 interacting zinc finger protein , resulted in
similar numbers of gamma-H2AX positive cells 1 h after
irradiation in comparison to the control transfected cells.
However, gamma-H2AX was more rapidly removed in these
cells (Figure 2B), potentially compromising effective DSBR. Taken
together, these results validate the effectiveness of our screen and
serve as an initial classification of molecular pathways for a
number of genes that can be explored in future studies.
KIAA0415 Is a Putative Helicase Required for Efficient HR-
For this work, we mined the screen by performing bioinfor-
matics analyses on the uncharacterized sequences in an attempt to
reveal possible molecular functions. KIAA0415 emerged as
particularly notable. By applying threading techniques (see Online
Methods for details), we identified potential structural homologies
of KIAA0415 with proteins belonging to the fold family ‘‘P-loop
containing nucleoside triphosphate hydrolases’’ (SCOP c.37;
Table S2). This fold family contains the so-called ‘‘helicase C
Gene NameEnsembl ID
Frequency of HR-DSBR
esiRNA 1esiRNA 2 CisplatinMMCIR
n.d n.dn.d n.d
n.d n.dn.d n.d
n.dn.d n.d n.d
n.d n.d n.dn.d
Calculated z-scores using esiRNA against Rluc as negative controls are shown for two independent esiRNA and marked with arrows. Genes that after knockdown
decreased the frequency of HR below z-score 24, 22, or 21.5 are marked with QQQ, QQ or Q, respectively; genes that increase the frequency of HR over z-score 4,
2, or 1.5 are marked with qqq, qq, or q, respectively; n.a., not available. Genes that decreased GFP levels with a z-score.4 are marked with +. Genes that after
knockdown decreased viability below 50% and 25% are marked with ++ and +, respectively. Genes that after knockdown decrease the cell number when treated with
cisplatin, MMC, or IR by 40%, 30%, or 10% in comparison to Rluc transfection are marked with +++, ++, and +, respectively; n.d., not done. In the last column arrows
indicate q increased and decreased Q number of gamma-H2AX positive cells after 1 h or 6 h post-IR for knockdowns statistically different (p,0.05) from Rluc
Table 1. Cont.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org5June 2010 | Volume 8 | Issue 6 | e1000408
domain’’ (PF00271) formed by a tandem repeat of two RecA-like
domains (Tandem AAA-ATPase superfamily). Top scoring
sequence-to-structurealignments were obtainedwith the
KIAA0415 sequence and the structure of the helicase C domains
of SF2 helicases that are involved in DNA repair such as UvrB,
Hel308, RecG, and TRCF (Figure 3). Visual inspection of the
gammaH2AX positive cells
no IR 1h recovery after IR6h recovery after IR
no treatment cisplatin MMC IR
Figure 2. Secondary assays for genes that decrease the frequency of HR-DSBR. (A) Effect on the cell viability after transfection of indicated
esiRNAs (black) in combination with cisplatin (dark grey), MMC (light grey), and IR (white) treatment are shown. Error bars indicate standard deviation.
All results were normalized to esiRNA against Rluc transfections. (B) Effect on percent of gammaH2AX positive cells after transfection of indicated
esiRNAs without irradiation (black), 1 h post-irradiation (dark grey), and 6 h post-irradiation (light grey). Error bars indicate standard deviation.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org6June 2010 | Volume 8 | Issue 6 | e1000408
generated 3D model (see Online Methods) confirmed the existence
of potential SF2 helicase motifs in KIAA0415 (Figure S1).
Molecular dynamics simulations were used to refine the
KIAA0415 model and corroborated its stability and its putative
ADP and Mg2+recognition (Video S1, Online Methods). These
results further support the prediction of a helicase-like domain
within KIAA0415 and substantiate the conservation in 3D of
residues important for its function as a putative SF2 helicase.
Based on these results, we decided to further elucidate possible
molecular functions of KIAA0415. We first tested the potency of
the employed KIAA0415 esiRNAs in more detail. Both esiRNAs
efficiently depleted KIAA0415 mRNA transcripts (Figure 4A) and
protein (Figure 4B). We then repeated the DR-GFP assay in the
HeLa reporter cell line and found 3.4 (esiRNA1) and 4.3
(esiRNA2) fold decrease in GFP positive cells in comparison to
controls, suggesting reduced frequencies of homologous recombi-
nation (Figure 4C). We examined the expression levels of I-SceI
after the knockdowns to rule out the possibility that I-SceI-
generated DSBs are compromised (Figure S2). To exclude a
possible cell-type specific effect, we also tested the knockdowns in a
different cell line. U2OS cells carrying a single insertion site of the
DR-GFP construct showed a similar reduction of GFP positive
cells upon KIAA0415 knockdown (Figure 4D), indicating that this
effect was not cell line specific. Finally, we excluded possible off-
target effects by performing cross-species RNAi rescue experi-
ments . Stable expression of mouse KIAA0415 in the human
DR-GFP cell line rendered this cell line resistant to the human
esiRNAs, authenticating a role of KIAA0415 in HR-DSBR
(Figure 4E). In summary, these results suggest that KIAA0415 is a
novel putative SF2 helicase required for efficient HR-DSBR.
KIAA0415 Forms a Complex with Proteins Associated
with Spastic Paraplegia
To further characterize KIAA0415, we tagged the gene on a
bacterial artificial chromosome (BAC) applying the TransgeneO-
mics approach . This method allows expression of tagged
proteins from its native promoter in its genomic context, and
therefore, the protein is expressed near physiological levels
[44,45]. C- and N-terminally tagged KIAA0415 was successfully
cloned and expressed in HeLa cells. The fusion protein showed
disperse, cytoplasmic, and nuclear localization, which did not
change considerably upon induction of DNA damage (unpub-
lished data). Immunoblotting of cell extracts revealed two major
protein bands, possibly reflecting two KIAA0415 isoforms
Figure 3. Structure-based sequence alignment of the helicase C domains of the SF2 helicases UvrB (2D7D), Hel308 (2P6R), RecG
(1GM5), and TRCF (2EYQ). Their consensus secondary structure elements are shown bellow as red spirals (a-helices) and blue arrows (b-strands).
The sequence alignment of KIAA0415 obtained from threading and used to build a 3D model of its putative helicase C-like domain based on these
structural templates is shown at the top. Sequence conservation of KIAA0415 with respect to the template structures is highlighted in grey
(conservative) and yellow (semi-conservative). Gap deletions and insertions are represented by dashed lines and inverted U symbols, respectively.
Insertions are labelled with the corresponding N- and C-ending residue numbering (black for KIAA0415, green for UvrB, and blue for Hel208). Regions
I, Ia, II, III, IV, and V of consensus SF2 helicase motifs are underlined. Residues involved in ADP- and Mg2+binding are coloured in blue and red,
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org7June 2010 | Volume 8 | Issue 6 | e1000408
(Figure 4B). Cell fractionations showed that the shorter isoform
was predominantly nuclear, whereas the longer form was found
mostly in the cytoplasm (Figure 4F). Immunoprecipitation
experiments followed by spectrometric identification of co-isolated
proteins revealed interactions of KIAA0415-LAP with SPG11,
SPG15, C20orf29, and DKFZp761E198 (Figure 5A,B and Table
S3). In order to validate these interactions we generated cell lines
expressing C-terminally tagged
DKFZp761E198 again using the TransgeneOmics approach.
Reciprocal immunoprecipitation experiments followed by mass
spectrometry analyses of in-gel and in-solution digests confirmed
the existence of a protein complex, which consists of at least five
core proteins: KIAA0415, SPG11, SPG15, C20orf29, and
DKFZp761E198 (Figure 5B and Table S3). In order to test
whether protein interaction partners of KIAA0415 would also
affect HR-DSBR, we tested esiRNAs targeting these genes in the
DR-GFP assay. Interestingly, significant reduction of GFP positive
cells were observed upon silencing of C20orf29 and SPG15 with
two independent esiRNAs (Figure 6), suggesting that these proteins
are also required for efficient HR-DSBR. Knockdown of SPG11
and DKFZp761E198, however, did no have an effect on the
percentage of GFP positive cells. Together, these experiments
reveal a novel protein complex, which at least in part is required
for efficient HR-DSBR.
KIAA0415 -1KIAA0415 -2
Figure 4. Functional analysis of KIAA0415. (A) Efficiency of KIAA0415 mRNA knockdown with two independent esiRNAs in HeLa cells. Relative
levels of mRNA 24 h post-transfection of indicated esiRNAs are shown. Error bars indicate standard deviation, **p,0.01. (B) Efficiency of KIAA0415
protein knockdown with two independent esiRNAs in HeLa cells. KIAA0415-LAP Western blot analysis of HeLa cell extracts 48 h post-transfection of
indicated esiRNAs is shown. A blot against tubulin served as the loading control. (C) Two independent KIAA0415 esiRNAs influence the frequency of
homologous recombination repair measured utilizing the DR-GFP assay. The relative percentage of GFP positive HeLa cells, normalized to the Rluc
transfected cells, is shown for indicated esiRNAs. Error bars indicate standard deviation. **p,0.01. (D) The KIAA0415 knockdown phenotype is not cell
line dependent. The relative percentage of GFP positive U2OS cells, normalized to the Rluc transfected cells, is shown for indicated esiRNAs. Error bars
indicate standard deviation, **p,0.01. (E) Expression of the mouse KIAA0415 orthologue rescues the KIAA0415 RNAi phenotype. The relative
percentage of GFP positive in DR-GFP HeLa cells stably expressing the mouse KIAA0415 from a BAC transfected with indicated esiRNAs are shown.
Error bars indicate standard deviation, **p,0.01. (F) Different KIAA0415 isoforms are found in the nucleus and in the cytoplasm. A KIAA0415 Western
blot of HeLa cell extracts after cell fractionation is shown. Blots against tubulin and histone H3 served as controls.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org8 June 2010 | Volume 8 | Issue 6 | e1000408
KIAA0415 Is Mutated in Patients with Spastic Paraplegia
The KIAA0415 interaction partners SPG11 and SPG15, also
known as spatacsin and spastizin, are encoded by two genes that
have been associated with hereditary spastic paraplegia with thin
corpus callosum (HSP-TCC) [46,47]. HSP-TCC is a subset of
hereditary spastic paraplegia (HSP), which are inherited neuro-
logical disorders caused by the degeneration of the cortico-spinal
tracts leading to lower-limb spasticity. HSP is a highly heteroge-
neous condition with at least 46 loci identified so far . A
potential interaction of SPG11 and SPG15 has been suggested on
the basis of similar neurological symptoms , however a
physical interaction of SPG11 and SPG15 has not been reported
thus far. Because of the physical interaction of KIAA0415 with
these two proteins encoded by genes associated with HSP, we
decided to investigate if any unexplained HSP cases could be
linked to mutations in KIAA0415. Direct sequencing of
KIAA0415 in 166 unrelated HSP patients, including 38 and 64
cases with a recessive or dominant inheritance pattern and 64
sporadic cases (see Online Methods), identified 7 known and 15
new variants, respectively. Most of these variants were not
considered causative, because they did not affect protein sequence,
were not predicted to alter correct splicing, or were also found
interaction partner or bait:
KIAA0415 SPG11 SPG15 C20orf29
Figure 5. KIAA0415 interacts with SPG11, SPG15, DKFZp761E198 and C20orf29. (A) SDS-PAGE gels obtained from the
immunoprecipitation of KIAA0415-LAP and SPG11-LAP. Baits (marked in green) and prey (marked in black) were identified by in-gel digestion
and nanoLC-MS/MS analysis (see Figure S3). Bands that are not marked represent unspecific background proteins or bait specific proteins (see Table
S3, and Online Methods). (B) The composition of KIAA0415 protein complex analyzed as established by shotgun-LC-MS/MS (see Table S3). The
number of matched detected peptides and protein sequence coverage are shown. Results for bait proteins are marked in bold.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org9June 2010 | Volume 8 | Issue 6 | e1000408
frequently in control samples (Table S4). However, one of these
identified variants led to a premature stop codon at position 527
(c.1413_1426del14/p.L471LfsX56, Table S4) and was absent in
158 Caucasian and 84 North-African control chromosomes. The
mutation was heterozygous and no other mutation or variant was
found in the coding sequence of KIAA0415 or in its regulatory
regions in this apparently sporadic patient (FSP-70-1). No other
subjects from the family were available for sampling and no copy
number variations were detected on chromosome 7 in the affected
patient (unpublished data), but small heterozygous rearrangements
or mutations in uncovered regions (unknown exons or introns)
might have escaped detection.
More interestingly, we also found a homozygous mutation in two
French siblings (FSP-083), which was not detected in 156 Caucasian
and 242 North-African control chromosomes. In these patients, a
complex indel in exon 2 (c.[80_83del4;79_84ins22], Figure 7C)
generates a frameshift and a stop codon following amino-acid 29
(p.R27LfsX3, Figure 7A). Interestingly, the insertion is an imperfect
quadruplication of the sequence CTGTAA(A), suggesting DNA
polymerase slippage during DNA synthesis as the mechanism for
introduction of the mutation. Both affected patients presented with
progressive spastic paraplegia associated with urinary incontinence
since age 50 and 49, respectively. Cerebral MRI was normal but
spinal hyperintensities at C3-C4 and C7 were observed in one. Both
parents died at the age of 72 and 77, respectively, of non neurological
causes. They originated from two neighbouring villages, but there
was no known consanguinity. However, the analysis of three close
microsatellite markers (D7S531, D7S517, and D7S1492) and the loss
of heterozygosity (LOH) search using CYTO_12 microarrays
(unpublished data) confirmed that the region is homozygous in both
affected patients (Figure 7B).
To further substantiate a role of KIAA0415 in DNA repair, we
compared drug sensitivity in lymphoblast cell lines established
from a patient carrying the KIAA0415 mutation (FSP-083-4) and
a patient carrying a mutation in SPG15 (FSP-708-22 ) to
control lymphoblast cell lines. Strikingly, the KIAA0415 mutant
cells were significantly (p,0.05) more sensitive to MMC and
bleomycin treatments compared to any of the control cell lines
(Figure 8). In addition, also the SPG15 cell line showed a mild
sensitivity to these drugs, phenocopying the results observed in
HeLa and U2OS cells.
Figure 6. KIAA0415 interactors are required for efficient HR-
DSBR. The relative percentage of GFP positive HeLa cells, normalized
to the Rluc transfected cells, is shown for indicated esiRNAs. Error bars
indicate standard deviation, *p,0.05, **p,0.01.
putative helicase fold
23456789 10 111213 141516 17
D7S517 (4.5 Mb)
patient 4patient 5
C C C T G C T G TA A A C T G TA A C T G TA A A C T G TA A A C T G
C C C G G AT
C T G TA A A C T G
exon 2: c.[80_83delGGAT; 79_84insTGCTGTAAACTGTAACTGTAAA]
Figure 7. KIAA0415 mutation in HSP patients. (A) Schematic representation of the KIAA0415 exon structure. The location of the homozygous
mutation and the predicted putative helicase fold (in green) are indicated. (B) Pedigree of KIAA0415 mutation in the FSP-083 family. Square symbols
represent men; the circles represent women. Symbols of dead subjects are crossed. The filled symbols indicate affected individuals. The numbers
below individuals are an internal reference for each sampled individual. Stars indicate sampled subjects. Segregation of the mutation and of 3
microsatellites markers on chromosome 7p are shown next to the pedigree, relative to their position in Mbases. The alleles of the microsatellites are
given in base pairs; M, mutation. (C) Electropherograms of KIAA0415 mutations. Deviating sequences from the wt sequence are underlined.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org 10June 2010 | Volume 8 | Issue 6 | e1000408
Taken together, these experiments identify KIAA0415 as a
novel gene, which is mutated in patients with HSP, and implicate
a link between HSP and DNA repair.
Using a well-characterized esiRNA library  we performed a
genome-scale RNAi screen and identified 61 genes that repro-
ducibly decreased or increased the frequency of DNA repair in an
assay for homologous recombination . Secondary assays for
processes relevant to DNA repair corroborated many of the initial
findings. Hence, we provide a dataset that should accelerate the
discovery of novel genes with roles in DNA repair and associated
medical conditions. Eighteen out of the 61 candidate genes have
been described in other large-scale mammalian DNA repair
studies [8,13,15,51], demonstrating the effectiveness of our screen,
but also highlighting that the use of different assays can uncover
novel players. Hence, we predict that the development of
alternative DNA repair assays for RNAi screens will reveal
additional genes implicated in DNA repair. For our screen we co-
transfected the ‘‘DNA damaging reagent,’’ I-SceI, together with
the esiRNA silencing triggers. Hence, proteins with long half-lives
may have been missed in this screen. Assays in which the DSB is
introduced some time after the cells were transfected with the
silencing triggers could uncover additional genes playing a role
during DNA repair.
To prioritize the molecular investigation of the uncharacterized
proteins identified in the screen, we employed a structural
bioinformatics approach. Based on the prediction that KIAA0415
represents a novel putative helicase we investigated this gene in
more detail. Tagging of the gene using the TransgeneOmics
approach revealed nuclear as well as cytoplasmic localization and
physical interaction with at least four proteins. Investigations of the
interaction partners showed that at least two of these proteins are
also required for efficient HR-DSBR. Possibly, these proteins form
a complex that is required for efficient HR-DSBR. Consequently,
the complex would lose its activity when one of the three proteins
Two of the interaction partners of KIAA0415 are encoded by
genes that are associated with spastic paraplegia. This result
prompted us to examine whether KIAA0415 mutations can
explain spasticity in patient samples not linked with mutations in
any of the known spastic paraplegia genes. We report a
homozygous mutation in KIAA0415, responsible for the spastic
paraplegia observed in two siblings. Hence, we identify KIAA0415
as a novel spastic paraplegia associated gene. Based on this finding,
we propose to rename KIAA0415 to SPG48 according to the
HUGO nomenclature. The fact that three proteins that form a
protein complex result in similar phenotypic consequences argues
that the whole complex is exerting an important function, which is
disturbed when one of the proteins is missing or non-functional. It
will therefore be interesting to investigate the remaining
interaction partners, C20orf29 and DKFZp761E198, for possible
mutations in HSP patients, even though they do not map to known
HSP loci . Although only demonstrated for one case, cell lines
derived from a patient carrying a SPG48 mutation were more
sensitive to DNA damaging drugs than control cells, corroborating
a role of SPG48 in DNA repair. Unfortunately, material from
other patients with SPG48 mutations was not available. However,
we propose that in the future HSP patients be screened for
mutations in SPG48 and that cells from these individuals be
checked for DNA repair defects.
Genes mutated in HSP have been associated with several
biological functions, including intracellular transport, axonal
pathfinding, mitochondrial functions, cholesterol metabolism,
myelin formation/stability, and chaperonin activity . Based
on our findings, we propose that HSP might also be a result of
impaired DNA repair, adding HSP to the growing list of
neurodegenerative diseases caused by DNA repair deficiencies
[4,5,7,8]. In agreement with this hypothesis, SPG11 has recently
been reported to be phosphorylated upon DNA damage by ATM
or ATR . Whether SPG48 (and its associated proteins) is a
PI negative cells
PI negative cells
day 0 day 1 day 2 day 3 day 4
Annexin V and
Annexin V and
Figure 8. Sensitivity of lymphoblast cell lines to DNA damaging
agents. (A) Growth curves of a control lymphoblast cell line—MHU-
2619 (solid line), a cell line carrying the KIAA0415 mutation—FSP-083-4
(dotted line) without drug treatment (black line) or MMC treatment
(light grey), are presented. (B and C) Cell lines derived from patients
with a mutation in KIAA0415 and SPG15 are more sensitive to DNA
damaging drugs. The decrease of viable cells (propidium iodide and
Annexin V negative) is shown in percent (grey bar) for the indicated cell
lines after MMC treatment (B) and bleomycin treatment (C). *p,0.05.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org 11June 2010 | Volume 8 | Issue 6 | e1000408
direct component of the HR-DSBR pathway or more indirectly
linked to DNA repair remains to be established. Biochemical
analysis of the putative helicase domain of SPG48 appears to be an
attractive entry point into gaining mechanistic insights into the
DNA repair function(s) of SPG48.
The technological advances in RNAi screening have increased
the speed at which phenotypic data can be obtained. However,
interpretation of the resulting genotype-phenotype relationships
remains challenging, and approaches that help to decipher the
screening data are highly desirable. Approaches that analyze
phenotypic data from unrelated RNAi screens and that combine
phenotypic- with localization- and proteomic data [52,53] have
been used successfully to bootstrap phenotype-to-function analy-
ses. Here, we explored the possibility of combining RNAi
screening data with structural bioinformatics approaches. The
obtained results demonstrate that this combination generates
valuable information, which helps to prioritize the follow-up
studies of uncharacterized candidate genes. We envision that the
design of an automatic pipeline to analyze possible structural and
functional features beyond protein sequence similarities will
further accelerate the characterization of genes identified in RNAi
screens. In the future, it will be important to combine the different
‘‘omics’’ and bioinformatics approaches to understand DNA
repair at a systems level and to further accelerate the discovery
of genes relevant to human pathology.
Materials and Methods
Generation of HeLa DR-GFP Cell Lines
Ten mg of the DR-GFP construct  were transfected into
2.56106HeLa cells using 12 ml Enhancer (Qiagen) and 14 ml
Effectene (Qiagen) according to the manufacturer’s protocol.
Stable cell lines were selected with 3 mg/ml puromycin (Sigma-
Aldrich) and single clones were obtained by FACS sorting on a
FACSAria (BD Biosciences). Colonies derived from individual
clones were expanded and tested for their behaviour after
transfection with a plasmid encoding the I-SceI endonuclease. A
cell line with virtually no GFP positive cells before I-SceI
treatment and high number of GFP positive cells after I-SceI
treatment was chosen for the screen.
Immunofluorescence Microscopy Analysis
Cells were grown on glass coverslips and fixed with 3%
paraformaldehyde (PFA) as described previously . Immuno-
fluorescence stainings were performed with a primary mouse anti-
GFP antibody (Roche Diagnostics, 1:4,000 dilution) and a
secondary donkey anti-mouse antibody conjugated to Alexa488
(Molecular Probes, 1:500 dilution). Genomic DNA was counter-
stained with ProLong Gold antifade reagent containing DAPI
(Invitrogen). Images were acquired on an Axioplan II Microscope
(Zeiss) operated through MetaMorph (Molecular Devices).
Western Blot Analysis
Western blot analysis was performed as described previously
. In this study the following primary antibodies were used:
mouse anti-GFP (Roche Diagnostics, 1:4,000 dilution), mouse
anti-DM1alpha tubulin (MPI-CBG Antibody Facility, 1:50,000
dilution), and rabbit anti-Histone H3 (Abcam 1:25,000 dilution).
Genome-Scale esiRNA Screen
The esiRNA library employed has been described elsewhere
[16,54]. For the screen the I-SceI expression plasmid  was co-
transfected with individual esiRNAs in an arrayed fashion. Briefly,
50 ng of each esiRNA in 5 ml TE Buffer was pipetted into 384-well
tissue culture plates (BD Biosciences) and stored at 220uC. Each
plate contained four esiRNAs against Rad51 as positive control (at
positions C3, C21, M5, M18) and 12 esiRNAs targeting renilla
luciferase (Rluc) as negative control (at positions C4, D3, D4, C22,
D21, D22, M6, N5, N6, M19, N18, N19 as shown in Figure 1C).
Using a multi-well dispenser (WellMate, Thermo Scientific) a
mixture of the I-SceI plasmid (12.75 ng/well) and the Enhancer
(0.142 ml/well) in 5 ml/well EC Buffer (Qiagen) was dispensed and
briefly spun in a Heraeus Multifuge 4KR (Thermo Electron
Corporation). After incubation for 5 min, Effectene (0.12 ml/well)
diluted in 5 ml/well EC Buffer was added to each well and plates
were briefly spun again. The transfection mixture was incubated
for 5 min to allow complex formation. In the meantime HeLa cells
carrying the DR-GFP reporter construct were harvested, counted,
and diluted to a final concentration of 40 cells/ml in DMEM
(Invitrogen) containing 12.5% Fetal Bovine Serum (Invitrogen).
Fifty ml of the cell suspension was added to each well using a multi-
well dispenser (Wellmate, Thermo Scientific). In order to prevent
evaporation, plates were sealed with breathable plate sealing foils
(Corning) and incubated in a tissue culture incubator at 37uC in
5% CO2. The medium was replaced 24 h post-transfection. After
another 72 h cells were washed with PBS and detached by adding
15 ml/well trypsin/EDTA (Invitrogen). After 25 min cells were
fixed by addition of 15 ml/well 3% PFA and stored no longer than
48 h at 4uC. Cells were assayed with a FACSCalibur (BD
Biosciences) equipped with a High Throughput Sampler (BD
Biosciences). Data were acquired and analyzed using CellQuest
Pro (BD Biosciences).
Z-scores were calculated for the percentages of GFP positive
cells using the following equation: z=(x2m) s21with: x 2
percentage of GFP positive cells; m 2 mean percentage of GFP
positive cells; s 2 standard deviation of the number of GFP
positive cells. In the primary screen mean and standard deviations
were calculated separately for each plate over all samples on the
plate excluding controls. Z-scores were calculated for each esiRNA
and averaged for duplicates. The transfection of esiRNA targeting
Rad51 was used as positive control and as reference for the assay
performance. esiRNAs for which the average z-score was below
22 or over 2 were considered as primary hits (Table S1).
In further validation experiments, the z-scores were calculated
based on the mean and standard deviation of the negative control
(Rluc transfection). EsiRNAs for which the average z-score for 4
replicates were below 22 or over 2 for one esiRNA and below
21.5 or over 1.5 for a second esiRNA were classified as validated
hits. Primer sequences for utilized esiRNAs are presented in Table
Gene enrichment analysis was performed using the Panther
Analysis Tools (http://www.pantherdb.org/tools/).
Cisplatin/MMC/IR Sensitivity Assay
Fifteen ng of each esiRNA diluted in 5 ml Opti-MEM
(Invitrogen) was pipetted in 384-well tissue culture plates (Greiner).
0.2 ml Oligofectamine (Invitrogen) was diluted with 4.8 ml Opti-
MEM, incubated for 5 min and pipetted to each well of the plate.
The mixtures were incubated for 20 min to allow complex
formation and 1,000 cells in 40 ml medium were added to each
well. Twenty-four hour post-transfection cisplatin (100 ng/ml) or
MMC (100 ng/ml) were added for 1 h or cells were exposed to
10 Gy IR. Cells were washed carefully with PBS and new medium
was added. After additional 48 h cells were fixed with 220uC
cold methanol for 20 min, washed twice with PBS, and blocked
with Blocking Buffer (0.2% Gelatin from cold water fish skin
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org 12June 2010 | Volume 8 | Issue 6 | e1000408
(Sigma-Aldrich Chemie) in PBS) for 5 min. Cell nuclei were
stained with DAPI (1 mg/ml) and cells were preserved with 0.02%
sodium azide in PBS. Images were acquired on an Olympus IX81
microscope (Olympus) and cell numbers were determined using
the Scan‘R Analysis software (Olympus). Every knockdown was
repeated 3 times. Cell numbers with and without DNA damaging
agents were compared to Rluc transfections.
HeLa cells were treated with 10 Gy IR 48 h post esiRNA
transfection and fixed 1 h or 6 h later. Cells were stained with a
phospho-H2AX antibody (clone JBW301, Upstate Biotechnology,
1:600 dilution) and with donkey anti-mouse TxRed conjugated
antibody (Molecular Probes, 1:400 dilution). DNA was stained
with DAPI (1 mg/ml). Cells were preserved with 0.02% sodium
azide in PBS and images were acquired on an Olympus IX81
microscope and analyzed by Scan‘R Analysis software (Olympus).
Every knockdown was repeated 3 times. Percentages of gamma-
H2AX positive cells were compared to Rluc transfections. p values
were calculated by Student’s t test.
Structural Bioinformatics Methods
Sequence-based analysis (Blast) failed to identify any statistically
significant sequence homology between KIAA0415 and any
previously characterized protein. Fold recognition techniques
were applied to search for potential structural homologies of
KIAA0415 with known protein structures. The threading
algorithm ProHit (ProCeryon Biosciences) was used to search for
structural resemblance of the uncharacterized KIAA0415 se-
quence with protein structures of the Brookhaven Protein
Databank (PDB). Threading calculations were performed with
parameters and scoring functions as previously published . A
fold library consisting of 19.961 protein chains representing the
PDB at 95% sequence identity was used. Three-dimensional (3D)
models for KIAA0415 were generated by threading its sequence
through each fold of the fold library. Inspection of fold coverage,
gaps position and content in the sequence-to-structure alignments
obtained, together with the analysis of the secondary structure
prediction obtained for KIAA0415 by PredictProtein (http://
www.predictprotein.org/) were used to discard possible false
positives in top scoring folds. A three-dimensional model of
KIAA0415 was built based on the threading alignments obtained
with high confidence predicted folds and four template structures
(PDBId: 2d7d, 2p6r, 1gm5, and 2eyq) by using Modeler in
Discovery Studio (Accelrys v1.7). Manual docking of ADP and
Mg2+onto the resulting KIAA0415 3D model was done based
on the X-ray structures of 2d7d and 1gm5. Refinement of the
obtained complex was done with AMBER 10 . A first step of
energy-minimization by 1,000 cycles of steepest descent and 500
cycles of conjugate gradient with harmonic force restraints on
protein atoms was followed by 3,000 cycles of steepest descent and
3,000 cycles of conjugate gradient without constraints. The system
was then heated from 0 to 300K for 10 ps. An equilibration step of
30 ps at 300K was followed by a 10 ns MD productive run. The
ff03 force field, periodic boundary conditions at constant pressure
with Langevin temperature coupling and Berendsen pressure
coupling, TIP3P explicit solvent, counterions, 8 A˚cut-off for non-
bonded interactions, and the SHAKE algorithm for hydrogens
BAC recombineering and the generation of BAC-transgenic cell
lines was performed as described previously [44,57]. A list of all
BACs and primers used in this study is provided in Table S6.
Immunoprecipitation and Mass Spectrometry Analysis
A goat anti-GFP antibody (MPI-CBG Antibody Facility)
immobilized on G-protein sepharose (GE Healthcare) or GFP-
Trap (Chromotek) were used for immunoprecipitation [44,52].
Glycine eluated KIAA0415-LAP and SPG11-LAP complexes
were analyzed on silver stained SDS PAGE. Excised slices were in-
gel digested and analyzed by nanoLC-MS/MS on a LTQ
(Thermo Fisher Scientific) as previously reported [58,59]. Glycine
eluates from KIAA0415-LAP, KIAA0415-NFLAP, SPG11-LAP,
and DKFZp761E198-LAP immunopurifications were used for in-
solution digestion and analyzed by shotgun-LC-MS/MS on a
LTQ Orbitrap (Thermo Fisher Scientific) . Proteins identified
in more than 15% of 193 independent immunoprecipitations
performed in ongoing collaborations projects from unrelated baits
were considered common backgrounds and further excluded.
Cell fractionation was performed with the commercially
available ProteoExtract kit (Novagene, Merck Biosciences) ac-
cording to the manufacturer’s protocol.
We selected 166 unrelated index cases with spastic paraplegia
diagnosed according to the Harding’s criteria ; 109 had a pure
form of the disease and 57 had a complex form partially
overlapping with the SPG11 typical phenotype. They included
64 index patients from families with dominant inheritance (mean
age at onset: 27.0616.6 y), 38 index patients with inheritance
compatible with an autosomal recessive trait (mean age at onset:
25.6619.9 y), and 64 patients with no family history of the disease
(mean age at onset: 31.2616.9 y). Most patients were French
(n=137) while the remaining patients originated from other
countries in Europe (n=16), North-Africa (n=8), or elsewhere
This study was approved by the local Bioethics committee
(approval No. 03-12-07 of the Comite ´ Consultatif pour la
Protection des Personnes et la Recherche Biome ´dicale Paris-
Necker to Drs A. Durr and A. Brice). Informed and written
consents were signed by all participating members of the families
before blood samples were collected for DNA extraction. All
clinical evaluations were performed according to a protocol
established by the European and Mediterranean network for
spinocerebellar degenerations (SPATAX, coordinator: Dr. A.
Durr) that included: a full medical history and examination,
estimation of the age at onset by the patient, observation of
additional neurological signs, electroneuromyographic (ENMG)
studies, and brain MRI, when possible. Disability was assessed on
a 7-point scale as previously described [62,63].
Mutations in SPAST, SPG3, SPG6, and SPG42 were
previously excluded in most of the index patients with dominant
transmission by direct sequencing and multiplex ligation-depen-
dent probe amplification in the case of SPAST and SPG3  and
unpublished data. Among autosomal recessive and sporadic
patients, mutations in the CYP7B1/SPG5 gene were excluded
in most patients  while SPG11 and SPG15 mutations have
been excluded in all complex autosomal recessive forms .
All coding exons of the gene KIAA0415 (Ensembl gene ID:
ENSG00000164917) and its splice junctions were amplified by
PCR on a Thermocycler 9700 (Applied Biosystems, Foster City,
CA) using specific primers (see Table S7). 3.1 Kb on the 39 and
1.5 Kb on the 59-UTRs were also sequenced in patients with an
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org 13June 2010 | Volume 8 | Issue 6 | e1000408
autosomal recessive transmission carrying a single heterozygote
variant. The amplicons were sequenced in both directions using
the BIGDYE V3 chemistry in an ABI3730 automated sequencer
(Applied Biosystems) as recommended by the supplier. The
seqscape v2.6 (Applied Biosystems) software was used to
highlight nucleotide variations in comparison to the normal
consensus sequence of both genes. In family FSP70, the
mutation was confirmed after subcloning of PCR products into
the pcDNA3.1/V5-His TOPO TA vector using TOP10
bacteria according to the manufacturer’s recommendations
(Invitrogen) and direct sequencing of at least 5 independent
clones of both alleles.
After identification of a variant, reamplification and resequencing
was systematically performed. Segregation of the mutations/
polymorphisms with the disease was verified by direct sequencing
in additional family members whose DNA samples were available. In
addition,79and121unrelatedhealthyCaucasian and North-African
subjects were screened to evaluate the frequency of new nucleotide
changes. In order to estimate evolutionary conservation, gene
sequences of different species were downloaded from the Ensembl
genome browser (www.ensembl.org) and aligned using the ClustalW
algorithm (http://www.ebi.ac.uk/Tools/clustalw2/index.html). All
variants were systematically tested for their effect on splicing
at: http://rulai.cshl.edu/cgi-bin/tools/ESE3/esefinder.cgi, http://
fruitfly.org/seq_tools/splice.html. Predicted effects of missense
changes were tested using SIFT and POLYPHEN at http://sift.
Lymphoblast Cell Lines
Cell lines were obtained from patients by infection with
Epstein-Barr-Virus (Table S8). Lymphoblast were cultured in
RPMI medium supplemented with 1% Pen/Strep, 2 mM L-
Glutamine, 10 mM Hepes, 1% Fungizone, and 20% FCS.
200.000 cells were plated in 6-well plates and cultured without or
with 10 ng/ml MMC or exposed to 10 ug/ml bleomycin for 1 h.
Growth of the cells was monitored daily by counting the trypan
blue negative cells using a Countess Automated Cell Counter
(Invitrogen). Four days after incubation 100.000 cells were
stained with the FITC Annexin V Appoptosis Kit II (BD
Biosciences) followed by FACS (BD Biosciences) analyses
following the manufacturer’s protocol. Experiments were per-
formed two times in duplicates.
plot of RMSD along the MD simulation (b).
Found at: doi:10.1371/journal.pbio.1000408.s001 (3.81 MB EPS)
3D model of the KIAA0415 SF2 domain (a) and
SceI expression levels.
Found at: doi:10.1371/journal.pbio.1000408.s002 (0.69 MB EPS)
KIAA0415 knockdown does not influence I-
Found at: doi:10.1371/journal.pbio.1000408.s003 (0.89 MB EPS)
Interacting proteins of the core KIAA0415
that decreased the frequency of homologous recombi-
nation below an average z-score of 2 22. Green, esiRNAs that
increased the frequency of homologous recombination above an
average z-score of 2. n.a., not available.
Found at: doi:10.1371/journal.pbio.1000408.s004 (3.90 MB XLS)
Primary RNAi screening data. Red, esiRNAs
Found at: doi:10.1371/journal.pbio.1000408.s005 (0.02 MB XLS)
shotgun-LC-MS/MS after immunoprecipitations. Pro-
teins were considered confident hits when identified with at least
three peptides with MASCOT ions score above 20.
Found at: doi:10.1371/journal.pbio.1000408.s006 (0.06 MB
Proteins identified by nanoLC-MS/MS and by
Found at: doi:10.1371/journal.pbio.1000408.s007 (0.03 MB XLS)
List of variants and mutations found in
esiRNAs. AR, autosomal recessive; AD, autosomal dominant;
spo, sporadic cases; ESE, Exonic splicing enhancers; na, not
applicable, dbSNP, single nucleotide polymorphisms database at
Found at: doi:10.1371/journal.pbio.1000408.s008 (0.04 MB XLS)
Primer sequences used to generate secondary
Found at: doi:10.1371/journal.pbio.1000408.s009 (0.02 MB XLS)
BAC clones and primer sequences used for
Found at: doi:10.1371/journal.pbio.1000408.s010 (0.02 MB XLS)
List of primers used for KIAA0415 mutation
Found at: doi:10.1371/journal.pbio.1000408.s011 (0.01 MB
Lymphoblast cell lines used for sensitivity
tive helicase C domain in KIAA0415 (see Online
Methods). The protein is shown in a cartoon representation.
The SF2 helicase motif regions are shown in colours: I in white, Ia
in yellow, II in orange, III in red, IV in cyan, and V in blue. ADP
and three residues (E379, D481, and E652) coordinating the Mg2+
are shown in sticks and coloured by atom type. Mg2+is
represented by a green sphere.
Found at: doi:10.1371/journal.pbio.1000408.s012 (1.56 MB
Molecular dynamics simulation of the puta-
We would like to thank D. Kappei, A. Sedello, V. Surendranath, R.
Kittler, and all members of the F. Buchholz laboratory for discussions and
F. Stewart for critical reading of the manuscript. We would like to
acknowledge M. Augsburg, M. Biesold, A. Ssykor, S. Rose, A. Weise, I.
Nu ¨sslein, R. Gey, L. Pingault, C. Tesson, J. Garrigues, and the High-
Throughput Technology Development Studio for technical assistance; M.
Jasin for the DR-GFP and I-SceI constructs and U2OS DR- GFP cell line;
K. Schumann for help with ionizing radiation experiments; and S. Forlani,
E. Martin, W. Carpentier, A. Rastetter, and the DNA and Cell Bank of the
Centre de Recherche de l’Institut du Cerveau et de la Moelle Epiniere for
their contribution to this study. We are grateful to the family members for
their participation. We also thank the clinicians who referred to us some of
the patients: D. Michel, B. Laurent, M. Abada-Bendib, and E. Ollagon-
The author(s) have made the following declarations about their
contributions: Conceived and designed the experiments: MS GS FB.
Performed the experiments: MS MT DBK SS EM MJ JT AKH JT MTP
GS. Analyzed the data: MS MPR MTP GS FB. Contributed reagents/
materials/analysis tools: IP FP CC CM AD JPC AB AS. Wrote the paper:
MS GS FB.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org 14 June 2010 | Volume 8 | Issue 6 | e1000408
References Download full-text
1. Hoeijmakers JH (2001) Genome maintenance mechanisms for preventing
cancer. Nature 411: 366–374.
2. Lombard DB, Chua KF, Mostoslavsky R, Franco S, Gostissa M, et al. (2005)
DNA repair, genome stability, and aging. Cell 120: 497–512.
3. Rass U, Ahel I, West SC (2007) Defective DNA repair and neurodegenerative
disease. Cell 130: 991–1004.
4. Savitsky K, Bar-Shira A, Gilad S, Rotman G, Ziv Y, et al. (1995) A single ataxia
telangiectasia gene with a product similar to PI-3 kinase. Science 268:
5. Stewart GS, Maser RS, Stankovic T, Bressan DA, Kaplan MI, et al. (1999) The
DNA double-strand break repair gene hMRE11 is mutated in individuals with
an ataxia-telangiectasia-like disorder. Cell 99: 577–587.
6. Faivre L, Le Merrer M, Lyonnet S, Plauchu H, Dagoneau N, et al. (2002)
Clinical and genetic heterogeneity of Seckel syndrome. Am J Med Genet 112:
7. Carney JP, Maser RS, Olivares H, Davis EM, Le Beau M, et al. (1998) The
hMre11/hRad50 protein complex and Nijmegen breakage syndrome: linkage of
double-strand break repair to the cellular DNA damage response. Cell 93:
8. Paulsen RD, Soni DV, Wollman R, Hahn AT, Yee MC, et al. (2009) A genome-
wide siRNA screen reveals diverse cellular processes and pathways that mediate
genome stability. Mol Cell 35: 228–239.
9. Wyman C, Kanaar R (2006) DNA double-strand break repair: all’s well that
ends well. Annu Rev Genet 40: 363–383.
10. Hartlerode AJ, Scully R (2009) Mechanisms of double-strand break repair in
somatic mammalian cells. Biochem J 423: 157–168.
11. Shrivastav M, De Haro LP, Nickoloff JA (2008) Regulation of DNA double-
strand break repair pathway choice. Cell Res 18: 134–147.
12. Lord CJ, McDonald S, Swift S, Turner NC, Ashworth A (2008) A high-
throughput RNA interference screen for DNA repair determinants of PARP
inhibitor sensitivity. DNA Repair (Amst) 7: 2010–2019.
13. Bartz SR, Zhang Z, Burchard J, Imakura M, Martin M, et al. (2006) Small
interfering RNA screens reveal enhanced cisplatin cytotoxicity in tumor cells
having both BRCA network and TP53 disruptions. Mol Cell Biol 26:
14. Doil C, Mailand N, Bekker-Jensen S, Menard P, Larsen DH, et al. (2009)
RNF168 binds and amplifies ubiquitin conjugates on damaged chromosomes to
allow accumulation of repair proteins. Cell 136: 435–446.
15. Stewart GS, Panier S, Townsend K, Al-Hakim AK, Kolas NK, et al. (2009) The
RIDDLE syndrome protein mediates a ubiquitin-dependent signaling cascade at
sites of DNA damage. Cell 136: 420–434.
16. Kittler R, Surendranath V, Heninger AK, Slabicki M, Theis M, et al. (2007)
Genome-wide resources of endoribonuclease-prepared short interfering RNAs
for specific loss-of-function studies. Nat Methods 4: 337–344.
17. Pierce AJ, Johnson RD, Thompson LH, Jasin M (1999) XRCC3 promotes
homology-directed repair of DNA damage in mammalian cells. Genes Dev 13:
18. West SC (2003) Molecular views of recombination proteins and their control.
Nat Rev Mol Cell Biol 4: 435–445.
19. Vasquez KM, Marburger K, Intody Z, Wilson JH (2001) Manipulating the
mammalian genome by homologous recombination. Proc Natl Acad Sci U S A
20. Lieber MR (2008) The mechanism of human nonhomologous DNA end joining.
J Biol Chem 283: 1–5.
21. Schorsch C, Kohler T, Boles E (2009) Knockout of the DNA ligase IV homolog
gene in the sphingoid base producing yeast Pichia ciferrii significantly increases
gene targeting efficiency. Curr Genet.
22. Bozas A, Beumer KJ, Trautman JK, Carroll D (2009) Genetic analysis of zinc-
finger nuclease-induced gene targeting in Drosophila. Genetics 182: 641–651.
23. Wood RD, Mitchell M, Sgouros J, Lindahl T (2001) Human DNA repair genes.
Science 291: 1284–1289.
24. Wood RD, Mitchell M, Lindahl T (2005) Human DNA repair genes, 2005.
Mutat Res 577: 275–283.
25. Tyagi S, Herr W (2009) E2F1 mediates DNA damage and apoptosis through
HCF-1 and the MLL family of histone methyltransferases. EMBO J 28:
26. Yang SZ, Lin FT, Lin WC (2008) MCPH1/BRIT1 cooperates with E2F1 in the
activation of checkpoint, DNA repair and apoptosis. EMBO Rep 9: 907–915.
27. Kobayashi K, Karran P, Oda S, Yanaga K (2005) Involvement of mismatch
repair in transcription-coupled nucleotide excision repair. Hum Cell 18:
28. Schurman SH, Hedayati M, Wang Z, Singh DK, Speina E, et al. (2009) Direct
and indirect roles of RECQL4 in modulating base excision repair capacity. Hum
Mol Genet 18: 3470–3483.
29. Laczmanska I, Gil J, Karpinski P, Stembalska A, Trusewicz A, et al. (2007)
Polymorphism in nucleotide excision repair gene XPC correlates with
bleomycin-induced chromosomal aberrations. Environ Mol Mutagen 48:
30. Petkovic M, Dietschy T, Freire R, Jiao R, Stagljar I (2005) The human
Rothmund-Thomson syndrome gene product, RECQL4, localizes to distinct
nuclear foci that coincide with proteins involved in the maintenance of genome
stability. J Cell Sci 118: 4261–4269.
31. Murakawa Y, Sonoda E, Barber LJ, Zeng W, Yokomori K, et al. (2007)
Inhibitors of the proteasome suppress homologous DNA recombination in
mammalian cells. Cancer Res 67: 8536–8543.
32. Gudmundsdottir K, Lord CJ, Ashworth A (2007) The proteasome is involved in
determining differential utilization of double-strand break repair pathways.
Oncogene 26: 7601–7606.
33. Gudmundsdottir K, Lord CJ, Witt E, Tutt AN, Ashworth A (2004) DSS1 is
required for RAD51 focus formation and genomic stability in mammalian cells.
EMBO Rep 5: 989–993.
34. Tomasz M (1995) Mitomycin-C - small, fast and deadly (but very selective).
Chemistry & Biology 2: 575–579.
35. Wang D, Lippard SJ (2005) Cellular processing of platinum anticancer drugs.
Nat Rev Drug Discov 4: 307–320.
36. Ward JF (1988) DNA damage produced by ionizing radiation in mammalian
cells: identities, mechanisms of formation, and reparability. Prog Nucleic Acid
Res Mol Biol 35: 95–125.
37. Sartori AA, Lukas C, Coates J, Mistrik M, Fu S, et al. (2007) Human CtIP
promotes DNA end resection. Nature 450: 509–514.
38. Burma S, Chen BP, Murphy M, Kurimasa A, Chen DJ (2001) ATM
phosphorylates histone H2AX in response to DNA double-strand breaks. J Biol
Chem 276: 42462–42467.
39. Ward IM, Chen J (2001) Histone H2AX is phosphorylated in an ATR-
dependent manner in response to replicational stress. J Biol Chem 276:
40. Keogh MC, Kim JA, Downey M, Fillingham J, Chowdhury D, et al. (2006) A
phosphatase complex that dephosphorylates gammaH2AX regulates DNA
damage checkpoint recovery. Nature 439: 497–501.
41. Hart MJ, Sharma S, elMasry N, Qiu RG, McCabe P, et al. (1996) Identification
of a novel guanine nucleotide exchange factor for the Rho GTPase. J Biol Chem
42. Wolf I, Rohrschneider LR (1999) Fiz1, a novel zinc finger protein interacting
with the receptor tyrosine kinase Flt3. J Biol Chem 274: 21478–21484.
43. Kittler R, Pelletier L, Ma C, Poser I, Fischer S, et al. (2005) RNA interference
rescue by bacterial artificial chromosome transgenesis in mammalian tissue
culture cells. Proc Natl Acad Sci U S A 102: 2396–2401.
44. Poser I, Sarov M, Hutchins JR, Heriche JK, Toyoda Y, et al. (2008) BAC
TransgeneOmics: a high-throughput method for exploration of protein function
in mammals. Nat Methods.
45. Bird AW, Hyman AA (2008) Building a spindle of the correct length in human
cells requires the interaction between TPX2 and Aurora A. J Cell Biol 182:
46. Stevanin G, Santorelli FM, Azzedine H, Coutinho P, Chomilier J, et al. (2007)
Mutations in SPG11, encoding spatacsin, are a major cause of spastic paraplegia
with thin corpus callosum. Nat Genet 39: 366–372.
47. Hanein S, Martin E, Boukhris A, Byrne P, Goizet C, et al. (2008) Identification
of the SPG15 gene, encoding spastizin, as a frequent cause of complicated
autosomal-recessive spastic paraplegia, including Kjellin syndrome. Am J Hum
Genet 82: 992–1002.
48. Stevanin G, Ruberg M, Brice A (2008) Recent advances in the genetics of spastic
paraplegias. Curr Neurol Neurosci Rep 8: 198–210.
49. Boukhris A, Stevanin G, Feki I, Denis E, Elleuch N, et al. (2008) Hereditary
spastic paraplegia with mental impairment and thin corpus callosum in Tunisia:
SPG11, SPG15, and further genetic heterogeneity. Arch Neurol 65: 393–402.
50. Goizet C, Boukhris A, Maltete D, Guyant-Marechal L, Truchetto J, et al. (2009)
SPG15 is the second most common cause of hereditary spastic paraplegia with
thin corpus callosum. Neurology 73: 1111–1119.
51. Matsuoka S, Ballif BA, Smogorzewska A, McDonald ER, 3rd, Hurov KE, et al.
(2007) ATM and ATR substrate analysis reveals extensive protein networks
responsive to DNA damage. Science 316: 1160–1166.
52. Theis M, Slabicki M, Junqueira M, Paszkowski-Rogacz M, Sontheimer J, et al.
(2009) Comparative profiling identifies C13orf3 as a component of the Ska
complex required for mammalian cell division. EMBO J 28: 1453–1465.
53. Hutchins JR, Toyoda Y, Hegemann B, Poser I, He ´riche ´ JK, et al. (2010)
Systematic analysis of human protein complexes identifies chromosome
segregation proteins. Science 328: 593–599.
54. Kittler R, Heninger AK, Franke K, Habermann B, Buchholz F (2005)
Production of endoribonuclease-prepared short interfering RNAs for gene
silencing in mammalian cells. Nat Methods 2: 779–784.
55. Pisabarro MT, Leung B, Kwong M, Corpuz R, Frantz GD, et al. (2006) Cutting
edge: novel human dendritic cell- and monocyte-attracting chemokine-like
protein identified by fold recognition methods. J Immunol 176: 2069–2073.
56. Case DA, Darden TA, Cheatham TEI, Simmerling CL, Wang J, et al. (2008)
AMBER 10. University of California, San Francisco.
57. Zhang Y, Buchholz F, Muyrers JP, Stewart AF (1998) A new logic for DNA
engineering using recombination in Escherichia coli. Nat Genet 20: 123–128.
58. Shevchenko A, Tomas H, Havlis J, Olsen JV, Mann M (2006) In-gel digestion
for mass spectrometric characterization of proteins and proteomes. Nat Protoc 1:
59. Shevchenko A, Roguev A, Schaft D, Buchanan L, Habermann B, et al. (2008)
Chromatin Central: towards the comparative proteome by accurate mapping of
the yeast proteomic environment. Genome Biol 9: R167.
DNA Repair RNAi Screen Identifies SPG48
PLoS Biology | www.plosbiology.org15June 2010 | Volume 8 | Issue 6 | e1000408